Let's see an example. The first thing we'll do is start up pandas and NumPy.
In the following screenshot, we have two series, srs1 and srs2:
srs1 has an index that goes from 0 to 4, whereas srs2 has an index that goes from 0 to 3, skips 4, and then goes to 5. These two series are technically the same length, but that doesn't necessarily mean that the elements will match up as you might expect. For example, let's consider the following code. What happens when we add srs1 and srs2?
Two NaNs were produced. That was because, for elements 0 to 3, there were elements in both series that could be matched up, but for 4 and 5, there were non-equivalent elements for each index in both series. This is also going to be the case when we multiply, shown as follows:
Or if we were to exponentiate, as follows:
That being said, Boolean arithmetic is different. In this case, comparison is done element by element, as you would normally expect. In fact, it seems that Boolean comparison doesn't care at...